
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from scipy import stats
import nltk
from wordcloud import WordCloud
from wordcloud import STOPWORDS, WordCloud
from PIL import Image
import plotly.express as px
import plotly.graph_objects as go


# Load  "Books data"
df_data=pd.read_csv('books_data-1.csv')


#Load  "Books rating"
df_rating = pd.read_csv('Books_rating.csv')


#Check the data
df_data.head(5)


#Check the data
df_rating.head(5)


## Books data
#Get a concise summary of the dataframe. 
df_data.info()
#Check for the number of null values for each variables.
df_data.isnull().sum()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 212404 entries, 0 to 212403
Data columns (total 10 columns):
 #   Column         Non-Null Count   Dtype  
---  ------         --------------   -----  
 0   Title          212403 non-null  object 
 1   description    143962 non-null  object 
 2   authors        180991 non-null  object 
 3   image          160329 non-null  object 
 4   previewLink    188568 non-null  object 
 5   publisher      136518 non-null  object 
 6   publishedDate  187099 non-null  object 
 7   infoLink       188568 non-null  object 
 8   categories     171205 non-null  object 
 9   ratingsCount   49752 non-null   float64
dtypes: float64(1), object(9)
memory usage: 16.2+ MB

Title                 1
description       68442
authors           31413
image             52075
previewLink       23836
publisher         75886
publishedDate     25305
infoLink          23836
categories        41199
ratingsCount     162652
dtype: int64


#### -Books Ratings
#Get a concise summary of the dataframe. 
df_rating.info()
#Check for the number of null values for each variables.
df_rating.isnull().sum()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3000000 entries, 0 to 2999999
Data columns (total 10 columns):
 #   Column              Dtype  
---  ------              -----  
 0   Id                  object 
 1   Title               object 
 2   Price               float64
 3   User_id             object 
 4   profileName         object 
 5   review/helpfulness  object 
 6   review/score        float64
 7   review/time         int64  
 8   review/summary      object 
 9   review/text         object 
dtypes: float64(2), int64(1), object(7)
memory usage: 228.9+ MB

Id                          0
Title                     208
Price                 2518829
User_id                561787
profileName            561886
review/helpfulness          0
review/score                0
review/time                 0
review/summary             38
review/text                 8
dtype: int64


#Delete variables that are irrelevant to the creation of the predictive model
df_data.drop(columns=['image', 'previewLink','infoLink','ratingsCount'], inplace=True)


df_data.head(5)


#Delete variables that are irrelevant to the creation of the predictive model
df_rating.drop(columns=['Id', 'User_id','review/helpfulness','profileName'], inplace=True)


df_rating.head(5)


#Do Margin-left Books data to Rating data via 'Title' column.
df=pd.merge(df_rating, df_data, on='Title', how='left')


df.head()


df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 3000000 entries, 0 to 2999999
Data columns (total 11 columns):
 #   Column          Dtype  
---  ------          -----  
 0   Title           object 
 1   Price           float64
 2   review/score    float64
 3   review/time     int64  
 4   review/summary  object 
 5   review/text     object 
 6   description     object 
 7   authors         object 
 8   publisher       object 
 9   publishedDate   object 
 10  categories      object 
dtypes: float64(2), int64(1), object(8)
memory usage: 274.7+ MB


# There were 2518829 missing values for 'Price' in 'Book ratings'. We replaced the average value of 'Price'.
print(df['Price'].mean())
#21.76265587495903

21.76265587495903


df['Price']=df['Price'].fillna(df['Price'].mean())
df.isnull().sum()
#The missing value for 'Price' is gone.

Title                208
Price                  0
review/score           0
review/time            0
review/summary        38
review/text            8
description       640225
authors           390634
publisher         782617
publishedDate     354581
categories        551498
dtype: int64


#The following graph shows the distribution of missing values in the data.
plt.figure(figsize=(10,10))
sns.heatmap(df.isnull(), cbar=False)

<AxesSubplot:>


#We removed all remaining missing values row by row. As a result, 2071176 data remained. 
#We determined that enough data remained to analyze.
df=df.dropna()
df.isnull().sum()

Title             0
Price             0
review/score      0
review/time       0
review/summary    0
review/text       0
description       0
authors           0
publisher         0
publishedDate     0
categories        0
dtype: int64


df.info()
df.shape

<class 'pandas.core.frame.DataFrame'>
Int64Index: 2071176 entries, 1 to 2999999
Data columns (total 11 columns):
 #   Column          Dtype  
---  ------          -----  
 0   Title           object 
 1   Price           float64
 2   review/score    float64
 3   review/time     int64  
 4   review/summary  object 
 5   review/text     object 
 6   description     object 
 7   authors         object 
 8   publisher       object 
 9   publishedDate   object 
 10  categories      object 
dtypes: float64(2), int64(1), object(8)
memory usage: 189.6+ MB

(2071176, 11)


#The following graph shows that there are no more missing values.
plt.figure(figsize=(10,10))
sns.heatmap(df.isnull(), cbar=False)

<AxesSubplot:>


####Found abnormal value in "review/time".


# seaborn boxplot:
plt.figure(figsize=(12,8))
sns.boxplot(x='review/score',y='review/time' ,data=df)
plt.title('review/time per review/score')
plt.xlabel('review/score')
plt.ylabel('review/time')
#The following graph shows the 'review/time per review/score'. The'review/score' includes Outlier at 1, 2 and 5, respectively.

Text(0, 0.5, 'review/time')


#We checked their Outlier and found that 'review/time' was "-1". This is an impossible value, so we removed it.
df[df["review/time"]<100]


#Remove to those datas
df=df[df['review/time'] != -1]


# seaborn boxplot:
plt.figure(figsize=(12,8))
sns.boxplot(x='review/score',y='review/time' ,data=df)
plt.title('review/time per review/score')
plt.xlabel('review/score')
plt.ylabel('review/time')
#The following graph shows that Outlier was successfully deleted.

Text(0, 0.5, 'review/time')


####Found abnormal value in "publishedDate".


#Create a new column named 'publishedDate_year' with only the year taken from 'publishedDate'.
df['publishedDate_year']=df['publishedDate'].str[:4]


#The "publishedDate_year" column had an unnatural value. ("19??", "199?","101-","2030")
#We Removed them.


df[df['publishedDate_year']=='19??']


df[df['publishedDate_year']=='199?']


df[df['publishedDate_year']=='101-']


df[df['publishedDate_year']=='2030']


df=df[df['publishedDate_year'] != '101-']


df=df[df['publishedDate_year'] != '2030']


df=df[df['publishedDate_year'] != '199?']


df=df[df['publishedDate_year'] != '19??']


#Changed from Object type to Int type
df['publishedDate_year'] = df['publishedDate_year'].astype(int)


df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 2071029 entries, 1 to 2999999
Data columns (total 12 columns):
 #   Column              Dtype  
---  ------              -----  
 0   Title               object 
 1   Price               float64
 2   review/score        float64
 3   review/time         int64  
 4   review/summary      object 
 5   review/text         object 
 6   description         object 
 7   authors             object 
 8   publisher           object 
 9   publishedDate       object 
 10  categories          object 
 11  publishedDate_year  int64  
dtypes: float64(2), int64(2), object(8)
memory usage: 205.4+ MB


#Randomly extracts 10000 pieces of data.
df_rating_sample10000=df.sample(n=10000)


#Create stopword list
stops= set(STOPWORDS)


#Making a Word cloud figure
textt = " ".join(x for x in df_rating_sample10000["review/text"])
wordcloud = WordCloud(stopwords=stops).generate(textt)
wordcloud.to_file("review.png")

<wordcloud.wordcloud.WordCloud at 0x7fedfe6bab80>


#showing a Word cloud figure
im = Image.open("review.png")
im.resize((900, 400))


#5 points is the highest frequency.
plt.figure(figsize=(12,8))
sns.countplot(x='review/score', data=df)

<AxesSubplot:xlabel='review/score', ylabel='count'>


#We wondered if there might be some causal relationship between the length of 'review/text' and the rating, 
#so we created a new 'text length' column that indicates the number of characters in each 'review/text'.
df["text length"] = df["review/text"].apply(len)
df.head()


#Let's examine the correlation between each 'review/score' and 'price', 'review/time' ,''publishedDate_year' and 'text length'.
score_gb = df.groupby('review/score').mean()
score_gb


# Showing the correlation heatmap 
sns.heatmap(df[["review/score","Price" ,"review/time","publishedDate_year","text length"]].corr(),cmap='coolwarm',annot=True)

<AxesSubplot:>


fig = go.Figure()
fig.add_trace(go.Histogram(x=df['publishedDate_year'], name='publishedDate_year'))
fig.update_traces(xbins=dict(start=1900,
                             end=2023,
                             size=0.5),
                  opacity=0.5)
fig.update_layout(barmode='overlay')
fig.show()

#After all, many of the books were published in the 2000s.


# In this study, a 'review/score' score of 4 or higher was defined as a high review, and a score of less than 4 was defined as a low review.
df['sentiment'] = ['positive' if x > 3 else 'negative' for x in df['review/score']]
df.head(3)


#Since there are about 2 million data, we randomly selected 10,000 each of 'Positive' and 'Negative'.


df_Psample10000=df[df['sentiment']=='positive'].sample(n=10000)


df_Nsample10000=df[df['sentiment']=='negative'].sample(n=10000)


#Create stopword list
stops= set(STOPWORDS)


#Removal of high frequency words regardless of high or low rating.
#Update the stopword list
stops.update(["book", "read","one","story","time","quot","character","books","good","make"])


###"review/text"


#Making a Word cloud figure(Positive)
textt_p = " ".join(x for x in df_Psample10000["review/text"])
wordcloud = WordCloud(stopwords=stops).generate(textt_p)
wordcloud.to_file("review_p.png")

<wordcloud.wordcloud.WordCloud at 0x7f907fce1fa0>


#showing a Word cloud figure
im = Image.open("review_p.png")
im.resize((900, 400))


#Making a Word cloud figure(Negative)
textt_n = " ".join(x for x in df_Nsample10000["review/text"])
wordcloud = WordCloud(stopwords=stops).generate(textt_n)
wordcloud.to_file("review_n.png")

<wordcloud.wordcloud.WordCloud at 0x7f907fce09a0>


#showing a Word cloud figure
im = Image.open("review_n.png")
im.resize((900, 400))


###"review/summary"


#Making a Word cloud figure(Positive)
textt_p = " ".join(x for x in df_Psample10000["review/summary"])
wordcloud = WordCloud(stopwords=stops).generate(textt_p)
wordcloud.to_file("review_summary_p.png")

<wordcloud.wordcloud.WordCloud at 0x7f9080bd9100>


#showing a Word cloud figure
im = Image.open("review_summary_p.png")
im.resize((900, 400))


#Making a Word cloud figure(Negative)
textt_n = " ".join(x for x in df_Nsample10000["review/summary"])
wordcloud = WordCloud(stopwords=stops).generate(textt_n)
wordcloud.to_file("review_summary_n.png")

<wordcloud.wordcloud.WordCloud at 0x7f9080bd9be0>


#showing a Word cloud figure
im = Image.open("review_summary_n.png")
im.resize((900, 400))


#df.to_csv("output_pd.csv", encoding="shift_jis")


#Execution of required modules.
import string
import re as re
import random as ran
import plotly
import matplotlib
import nltk as nltk
from nltk.corpus import stopwords
from sklearn.linear_model import LogisticRegression
from sklearn.feature_extraction.text import CountVectorizer

pd.set_option('display.max_columns', None)
pd.set_option('display.max_rows', None)


df.head()


#Extract only the 'review/score', 'review/text', and 'sentiment' columns used for training.
df_sub = df[['review/score','review/text',  'sentiment']]
df_sub.head()


#Test data cleansing step:
#Write a function to remove all the punctuations and stop words from a given text.
#It will also remove duplicate words of the input text.  
#The function will accept one str argument and return one str value. 

def remove_punc_stopwords(text):
    text = re.sub(f"[{string.punctuation}]"," ",text)
    text_tokens = set(text.split())
    text_tokens = text_tokens.difference(stops)
    return " ".join(text_tokens)


#We randomly extracted 100000 pieces of data.
df_sub_sample100000=df_sub.sample(n=100000)


#Remove punctuations and stopwords from the text data 
df_sub_sample100000['review/text'] = df_sub_sample100000['review/text'].apply(remove_punc_stopwords)


df_sub_sample100000.head()


# split the dataset into two: train (85% of the obs.) and test (15% of the obs.)
df_sub_sample100000['random_index'] =[ ran.uniform(0,1) for x in range(len(df_sub_sample100000))]


df_sub_train=df_sub_sample100000[df_sub_sample100000.random_index < 0.85]
df_sub_test=df_sub_sample100000[df_sub_sample100000.random_index >= 0.85]


#count vectorizer:
#Next, we will use a count vectorizer from the Scikit-learn library. This will transform the text in our dataframe into a 'bag of words' model, which will contain a sparse matrix of integers. 
#The number of occurrences of each word will be counted and printed.
#We will need to convert the text into a bag-of-words model since the logistic regression algorithm cannot understand text.

print(df_sub_train.head())
print(df_sub_test.head())

vectorizer = CountVectorizer(token_pattern=r'\b\w+\b')
train_matrix = vectorizer.fit_transform(df_sub_train['review/text'])
test_matrix = vectorizer.transform(df_sub_test['review/text'])

         review/score                                        review/text  \
1578389           5.0  wellbeing figure even noggin wisdoms heartedly...   
1840683           3.0  suggest gone past seemed blood I might descrip...   
1441591           2.0  stands 1 18 humility doesn Taoist truths perso...   
2564758           3.0  practically pretty paragraph Robards total AND...   
956313            1.0  days Einstein evolutionists confess makes appl...   

        sentiment  random_index  
1578389  positive      0.379231  
1840683  negative      0.668277  
1441591  negative      0.056891  
2564758  negative      0.279974  
956313   negative      0.659383  
         review/score                                        review/text  \
758886            4.0  give So far draw don explain high written pret...   
37218             5.0  days disclosure far particular written third o...   
259778            5.0  complicated involvement written Though might f...   
1267691           5.0  notes s price 3 comments whatnot Nietzche edit...   
1684936           4.0  promises order favortie rewards power somethin...   

        sentiment  random_index  
758886   positive      0.974274  
37218    positive      0.896959  
259778   positive      0.941015  
1267691  positive      0.983960  
1684936  positive      0.934039


#Perform Logistic Regression
lr = LogisticRegression(solver='liblinear')
X_train = train_matrix
X_test = test_matrix
y_train = df_sub_train['sentiment']
y_test = df_sub_test['sentiment']

lr.fit(X_train,y_train)

LogisticRegression(solver='liblinear')


#Generate the predictions for the test dataset
predictions = lr.predict(X_test)
df_sub_test['predictions'] = predictions
print(df_sub_test.head(30))

         review/score                                        review/text  \
758886            4.0  give So far draw don explain high written pret...   
37218             5.0  days disclosure far particular written third o...   
259778            5.0  complicated involvement written Though might f...   
1267691           5.0  notes s price 3 comments whatnot Nietzche edit...   
1684936           4.0  promises order favortie rewards power somethin...   
2811160           5.0  due studies readers written difficulties rathe...   
608152            5.0  ring furthermore written twice series think no...   
2656286           5.0  Samsa great I twice short cause think assignme...   
496726            5.0  sets affects written makes cosmic Left terms p...   
1783028           4.0  stay far Burns life hero relentlessly sought I...   
540931            5.0  Also England check detailed Willis Watch short...   
1420872           5.0  great I The love movie Gweneth wonderful chara...   
916848            5.0  subject sweet Short concise I informative with...   
915653            5.0  past small fast money The stand lot still broo...   
2916243           5.0  reservation ultimately None written spirit ple...   
944661            5.0  tire Braun lifted series several Ms never worl...   
576772            4.0  30 Giacomin series fan case short even string ...   
126851            4.0  dig end someone will grave may careful Be want...   
2103549           2.0  detested vertitable repititious length ready o...   
2999930           4.0  sets police Zudenigo heinous nut happens colum...   
2049774           5.0  romantic great life girls As stories love judg...   
1380112           4.0  gave media characters publication She I NY dau...   
1527646           5.0  10 spring think mad 17 drinking One wants adve...   
2131178           5.0  fishies Target far Wood pretty makes still hum...   
1307815           1.0  days Not rapes great serious trust I gets host...   
1009641           5.0  Timothy entries told teach Keats opened Eighte...   
1401891           5.0  far I want absolutly 10 best things C 000 top ...   
1776680           5.0  feel sent threatened intellectuallism better p...   
1013050           5.0  device written Bradbury language 40s assumptio...   
1325354           5.0  teenagers written The difficult work crime imp...   

        sentiment  random_index predictions  
758886   positive      0.974274    positive  
37218    positive      0.896959    positive  
259778   positive      0.941015    positive  
1267691  positive      0.983960    negative  
1684936  positive      0.934039    positive  
2811160  positive      0.878055    positive  
608152   positive      0.926718    positive  
2656286  positive      0.864898    positive  
496726   positive      0.922799    positive  
1783028  positive      0.966678    positive  
540931   positive      0.987269    positive  
1420872  positive      0.881157    positive  
916848   positive      0.986534    positive  
915653   positive      0.931477    positive  
2916243  positive      0.948920    positive  
944661   positive      0.865909    positive  
576772   positive      0.869454    positive  
126851   positive      0.882837    positive  
2103549  negative      0.946415    negative  
2999930  positive      0.899608    positive  
2049774  positive      0.935930    positive  
1380112  positive      0.969562    positive  
1527646  positive      0.872566    positive  
2131178  positive      0.902218    positive  
1307815  negative      0.967857    negative  
1009641  positive      0.881208    positive  
1401891  positive      0.995687    positive  
1776680  positive      0.872125    positive  
1013050  positive      0.909471    positive  
1325354  positive      0.933354    negative

<ipython-input-84-cef1e152ef50>:3: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy


#Calculate the prediction accuracy
df_sub_test['match'] = df_sub_test['sentiment'] == df_sub_test['predictions']
print(sum(df_sub_test['match'])/len(df_sub_test))
#We created a good model with about 87% accuracy.

0.8685206079511515

<ipython-input-85-a917b4a5cf2d>:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy


from sklearn.metrics import confusion_matrix,classification_report


print(confusion_matrix(y_test,predictions))
print('\n')
print(classification_report(y_test,predictions))

[[ 1833  1247]
 [  734 11253]]


              precision    recall  f1-score   support

    negative       0.71      0.60      0.65      3080
    positive       0.90      0.94      0.92     11987

    accuracy                           0.87     15067
   macro avg       0.81      0.77      0.78     15067
weighted avg       0.86      0.87      0.86     15067


#Perform Logistic Regression
lr2 = LogisticRegression(solver='liblinear')
X_train = train_matrix
X_test = test_matrix
y_train2 = df_sub_train['review/score']
y_test2 = df_sub_test['review/score']

lr2.fit(X_train,y_train2)

LogisticRegression(solver='liblinear')


#Generate the predictions for the test dataset
predictions2 = lr2.predict(X_test)
df_sub_test['predictions_rating'] = predictions2
print(df_sub_test.head(30))

         review/score                                        review/text  \
758886            4.0  give So far draw don explain high written pret...   
37218             5.0  days disclosure far particular written third o...   
259778            5.0  complicated involvement written Though might f...   
1267691           5.0  notes s price 3 comments whatnot Nietzche edit...   
1684936           4.0  promises order favortie rewards power somethin...   
2811160           5.0  due studies readers written difficulties rathe...   
608152            5.0  ring furthermore written twice series think no...   
2656286           5.0  Samsa great I twice short cause think assignme...   
496726            5.0  sets affects written makes cosmic Left terms p...   
1783028           4.0  stay far Burns life hero relentlessly sought I...   
540931            5.0  Also England check detailed Willis Watch short...   
1420872           5.0  great I The love movie Gweneth wonderful chara...   
916848            5.0  subject sweet Short concise I informative with...   
915653            5.0  past small fast money The stand lot still broo...   
2916243           5.0  reservation ultimately None written spirit ple...   
944661            5.0  tire Braun lifted series several Ms never worl...   
576772            4.0  30 Giacomin series fan case short even string ...   
126851            4.0  dig end someone will grave may careful Be want...   
2103549           2.0  detested vertitable repititious length ready o...   
2999930           4.0  sets police Zudenigo heinous nut happens colum...   
2049774           5.0  romantic great life girls As stories love judg...   
1380112           4.0  gave media characters publication She I NY dau...   
1527646           5.0  10 spring think mad 17 drinking One wants adve...   
2131178           5.0  fishies Target far Wood pretty makes still hum...   
1307815           1.0  days Not rapes great serious trust I gets host...   
1009641           5.0  Timothy entries told teach Keats opened Eighte...   
1401891           5.0  far I want absolutly 10 best things C 000 top ...   
1776680           5.0  feel sent threatened intellectuallism better p...   
1013050           5.0  device written Bradbury language 40s assumptio...   
1325354           5.0  teenagers written The difficult work crime imp...   

        sentiment  random_index predictions  match  predictions_rating  
758886   positive      0.974274    positive   True                 3.0  
37218    positive      0.896959    positive   True                 5.0  
259778   positive      0.941015    positive   True                 4.0  
1267691  positive      0.983960    negative  False                 5.0  
1684936  positive      0.934039    positive   True                 5.0  
2811160  positive      0.878055    positive   True                 5.0  
608152   positive      0.926718    positive   True                 5.0  
2656286  positive      0.864898    positive   True                 5.0  
496726   positive      0.922799    positive   True                 5.0  
1783028  positive      0.966678    positive   True                 4.0  
540931   positive      0.987269    positive   True                 5.0  
1420872  positive      0.881157    positive   True                 5.0  
916848   positive      0.986534    positive   True                 5.0  
915653   positive      0.931477    positive   True                 4.0  
2916243  positive      0.948920    positive   True                 5.0  
944661   positive      0.865909    positive   True                 5.0  
576772   positive      0.869454    positive   True                 5.0  
126851   positive      0.882837    positive   True                 5.0  
2103549  negative      0.946415    negative   True                 1.0  
2999930  positive      0.899608    positive   True                 4.0  
2049774  positive      0.935930    positive   True                 5.0  
1380112  positive      0.969562    positive   True                 5.0  
1527646  positive      0.872566    positive   True                 5.0  
2131178  positive      0.902218    positive   True                 5.0  
1307815  negative      0.967857    negative   True                 5.0  
1009641  positive      0.881208    positive   True                 5.0  
1401891  positive      0.995687    positive   True                 5.0  
1776680  positive      0.872125    positive   True                 5.0  
1013050  positive      0.909471    positive   True                 5.0  
1325354  positive      0.933354    negative  False                 1.0

<ipython-input-89-c163005da33a>:3: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy


#Calculate the prediction accuracy
df_sub_test['match_rating'] = df_sub_test['review/score'] == df_sub_test['predictions_rating']
print(sum(df_sub_test['match_rating'])/len(df_sub_test))
#The 'Rating' prediction on a 5-point scale was still less accurate. #65%

0.6465786155173558

<ipython-input-93-dd2cd0dd3206>:2: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy


print(confusion_matrix(y_test2,predictions2))
print('\n')
print(classification_report(y_test2,predictions2))

[[ 474   99   79   71  282]
 [ 177  130  122   98  222]
 [  58   97  363  348  460]
 [  50   41  220  846 1795]
 [  71   44  156  835 7929]]


              precision    recall  f1-score   support

         1.0       0.57      0.47      0.52      1005
         2.0       0.32      0.17      0.22       749
         3.0       0.39      0.27      0.32      1326
         4.0       0.38      0.29      0.33      2952
         5.0       0.74      0.88      0.80      9035

    accuracy                           0.65     15067
   macro avg       0.48      0.42      0.44     15067
weighted avg       0.61      0.65      0.62     15067

	Title	description	authors	publisher	publishedDate	categories
0	Its Only Art If Its Well Hung!	NaN	['Julie Strain']	NaN	1996	['Comics & Graphic Novels']
1	Dr. Seuss: American Icon	Philip Nel takes a fascinating look into the k...	['Philip Nel']	A&C Black	2005-01-01	['Biography & Autobiography']
2	Wonderful Worship in Smaller Churches	This resource includes twelve principles in un...	['David R. Ray']	NaN	2000	['Religion']
3	Whispers of the Wicked Saints	Julia Thomas finds her life spinning out of co...	['Veronica Haddon']	iUniverse	2005-02	['Fiction']
4	Nation Dance: Religion, Identity and Cultural ...	NaN	['Edward Long']	NaN	2003-03-01	NaN

	Id	Title	Price	User_id	profileName	review/helpfulness	review/score	review/time	review/summary	review/text
0	1882931173	Its Only Art If Its Well Hung!	NaN	AVCGYZL8FQQTD	Jim of Oz "jim-of-oz"	7/7	4.0	940636800	Nice collection of Julie Strain images	This is only for Julie Strain fans. It's a col...
1	0826414346	Dr. Seuss: American Icon	NaN	A30TK6U7DNS82R	Kevin Killian	10/10	5.0	1095724800	Really Enjoyed It	I don't care much for Dr. Seuss but after read...
2	0826414346	Dr. Seuss: American Icon	NaN	A3UH4UZ4RSVO82	John Granger	10/11	5.0	1078790400	Essential for every personal and Public Library	If people become the books they read and if "t...
3	0826414346	Dr. Seuss: American Icon	NaN	A2MVUWT453QH61	Roy E. Perry "amateur philosopher"	7/7	4.0	1090713600	Phlip Nel gives silly Seuss a serious treatment	Theodore Seuss Geisel (1904-1991), aka "D...
4	0826414346	Dr. Seuss: American Icon	NaN	A22X4XUPKF66MR	D. H. Richards "ninthwavestore"	3/3	4.0	1107993600	Good academic overview	Philip Nel - Dr. Seuss: American IconThis is b...

	Title	description	authors	publisher	publishedDate	categories
0	Its Only Art If Its Well Hung!	NaN	['Julie Strain']	NaN	1996	['Comics & Graphic Novels']
1	Dr. Seuss: American Icon	Philip Nel takes a fascinating look into the k...	['Philip Nel']	A&C Black	2005-01-01	['Biography & Autobiography']
2	Wonderful Worship in Smaller Churches	This resource includes twelve principles in un...	['David R. Ray']	NaN	2000	['Religion']
3	Whispers of the Wicked Saints	Julia Thomas finds her life spinning out of co...	['Veronica Haddon']	iUniverse	2005-02	['Fiction']
4	Nation Dance: Religion, Identity and Cultural ...	NaN	['Edward Long']	NaN	2003-03-01	NaN

	Title	Price	review/score	review/time	review/summary	review/text
0	Its Only Art If Its Well Hung!	NaN	4.0	940636800	Nice collection of Julie Strain images	This is only for Julie Strain fans. It's a col...
1	Dr. Seuss: American Icon	NaN	5.0	1095724800	Really Enjoyed It	I don't care much for Dr. Seuss but after read...
2	Dr. Seuss: American Icon	NaN	5.0	1078790400	Essential for every personal and Public Library	If people become the books they read and if "t...
3	Dr. Seuss: American Icon	NaN	4.0	1090713600	Phlip Nel gives silly Seuss a serious treatment	Theodore Seuss Geisel (1904-1991), aka "D...
4	Dr. Seuss: American Icon	NaN	4.0	1107993600	Good academic overview	Philip Nel - Dr. Seuss: American IconThis is b...

	Title	Price	review/score	review/time	review/summary	review/text	description	authors	publisher	publishedDate	categories
0	Its Only Art If Its Well Hung!	NaN	4.0	940636800	Nice collection of Julie Strain images	This is only for Julie Strain fans. It's a col...	NaN	['Julie Strain']	NaN	1996	['Comics & Graphic Novels']
1	Dr. Seuss: American Icon	NaN	5.0	1095724800	Really Enjoyed It	I don't care much for Dr. Seuss but after read...	Philip Nel takes a fascinating look into the k...	['Philip Nel']	A&C Black	2005-01-01	['Biography & Autobiography']
2	Dr. Seuss: American Icon	NaN	5.0	1078790400	Essential for every personal and Public Library	If people become the books they read and if "t...	Philip Nel takes a fascinating look into the k...	['Philip Nel']	A&C Black	2005-01-01	['Biography & Autobiography']
3	Dr. Seuss: American Icon	NaN	4.0	1090713600	Phlip Nel gives silly Seuss a serious treatment	Theodore Seuss Geisel (1904-1991), aka "D...	Philip Nel takes a fascinating look into the k...	['Philip Nel']	A&C Black	2005-01-01	['Biography & Autobiography']
4	Dr. Seuss: American Icon	NaN	4.0	1107993600	Good academic overview	Philip Nel - Dr. Seuss: American IconThis is b...	Philip Nel takes a fascinating look into the k...	['Philip Nel']	A&C Black	2005-01-01	['Biography & Autobiography']

Goal: Demonstrate proficiency in text mining (sentiment analysis) using Python.¶

Synopsis:¶

The data package contains two files each corresponding to 212404 unique books that were available on Amazon between 1996 and 2014. 142.8 million reviews spanning May 1996 - July 2014 have also been collected. The two data files maintain the following structure:¶

Execution of required modules.¶

¶

¶

Task 2: Data Cleansing, Classification, and Word Clouds .¶

1. [5 Marks] Load the data into Python and perform the initial sanity checks.¶

¶

1-1. Variable Analysis¶

Check the structure of those data and the number of null values.¶

¶

1-2. Data Merge¶

¶

1-3. Treating missing values¶

¶

1-4. Treating outlier¶

¶

2. [10 Marks] Generate initial analytical visualizations to understand the reviews (i.e., histogram) including a word cloud (using the wordcloud package). Discuss your findings.¶

2-1. Visualize what words are often included in 'review/text' using a word cloud.¶

2-2.　Display of number of data per review score¶

2-3. A number of characters in each 'review/text'.¶

¶

2-4. Visualize the number of data per publication year.¶

¶

3. [5 Marks] Classify reviews into two ‘sentiment’ categories called positive and negative.¶

¶

4. [10 Marks] Generate positive and negative word clouds. Discuss your findings while comparing the positive and negative summaries (you may include other graphs if needed).¶

Using Word Cloud, we visualized the high-frequency WORDS for each high and low review.¶

Task 3: Prediction¶

1. [15 Marks] Build a simple logistic regression model to predict the sentiment category based on a text- based review. Discuss your findings.¶

¶

2. [15 Marks] Build a multinomial logistic regression model to predict the rating of a book based on its text-based review. Discuss your findings.¶

Fin.¶

	Title	Price	review/score	review/time	review/summary	review/text	description	authors	publisher	publishedDate	categories
75745	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	2.0	-1	For once, the movie was better.	I purchased this book after seeing, and truly ...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75746	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	1.0	-1	Dazed and Confused	I had such high hopes for this book and I was ...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75747	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	2.0	-1	Disappointing...read My Life in France by Juli...	I eagerly snatched this book up when I saw it ...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75748	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	2.0	-1	Narcissistic Hipsters Should Cook, Not Write	After seeing the movie, I knew what I was in f...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75749	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	1.0	-1	"Taste"-less	Poor Julia Child, she deserved a better tribut...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75750	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	1.0	-1	Meanspirited woman	The writing was okayish... But these details f...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75751	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	1.0	-1	Leaves a bitter taste...	Julie and Julia looked so charming, and starte...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75752	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	2.0	-1	Leave out the swearing please!!!!!	This book would have been ten times better (an...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75753	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	1.0	-1	waste of time and money	A book as messy and nasty as some of the autho...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
75754	Julie and Julia: 365 Days, 524 Recipes, 1 Tiny...	21.762656	2.0	-1	disappointing	I love both cooking and reading blogs, so I th...	Pushing thirty, living in a rundown apartment ...	['Julie Powell']	Viking	2005	['Cookery, French']
2152018	McKeachie's Teaching Tips, Strategies Research...	21.762656	5.0	-1	Pretty good.	Pretty good at providing me with the fundament...	From the Publisher: This indispensable handboo...	['Wilbert James McKeachie', 'Marilla Svinicki'...	Wadsworth Publishing Company	2006	['Education']

	Title	Price	review/score	review/time	review/summary	review/text	description	authors	publisher	publishedDate	categories	publishedDate_year
296184	Within a budding grove,	21.762656	3.0	1265414400	Say Hello to the Princess of Luxembourg	Proust is an odd one -- an essayist and memoir...	“A l’ombre des jeunes fi lles en fl eurs” est ...	['Proust M.']	Рипол Классик	19??	['Fiction']	19??
296185	Within a budding grove,	21.762656	5.0	978566400	Stunning!	Without a doubt, this is the best piece of lit...	“A l’ombre des jeunes fi lles en fl eurs” est ...	['Proust M.']	Рипол Классик	19??	['Fiction']	19??
296186	Within a budding grove,	21.762656	5.0	1330128000	Seascape with a frieze of boys...	Reading Proust is like drinking good tea. You ...	“A l’ombre des jeunes fi lles en fl eurs” est ...	['Proust M.']	Рипол Классик	19??	['Fiction']	19??
296187	Within a budding grove,	21.762656	5.0	1239840000	good edition of a classic	Though Proust is not for the casual reader, th...	“A l’ombre des jeunes fi lles en fl eurs” est ...	['Proust M.']	Рипол Классик	19??	['Fiction']	19??
296188	Within a budding grove,	21.762656	5.0	1215820800	Philosophy as narrative	Volume two of Le Proust's great work is a sens...	“A l’ombre des jeunes fi lles en fl eurs” est ...	['Proust M.']	Рипол Классик	19??	['Fiction']	19??
...	...	...	...	...	...	...	...	...	...	...	...	...
2655929	Melodious Etudes for Trombone Book 1	21.762656	5.0	916272000	Outstanding Teaching of the "Basics"!!!	This book teaches basic skills in which any st...	120 Melodious Etudes for Trombone, Book 1. Fro...	['Joannes Rochut']	Ravenio Books	19??	['Music']	19??
2655930	Melodious Etudes for Trombone Book 1	21.762656	5.0	1340150400	music book	Book is great! Just what my son needed. The pa...	120 Melodious Etudes for Trombone, Book 1. Fro...	['Joannes Rochut']	Ravenio Books	19??	['Music']	19??
2655931	Melodious Etudes for Trombone Book 1	21.762656	5.0	1318982400	O1594 - Melodious Etudes for Trombone Book 1 (...	This is a great book to work out of, I love it...	120 Melodious Etudes for Trombone, Book 1. Fro...	['Joannes Rochut']	Ravenio Books	19??	['Music']	19??
2655932	Melodious Etudes for Trombone Book 1	21.762656	5.0	1293494400	Best Legato Studies Ever	This Book will help you develop the technique ...	120 Melodious Etudes for Trombone, Book 1. Fro...	['Joannes Rochut']	Ravenio Books	19??	['Music']	19??
2655933	Melodious Etudes for Trombone Book 1	21.762656	5.0	1286323200	Melodious Etudes for Trombone Book 1	This is an excellent, well written skill book....	120 Melodious Etudes for Trombone, Book 1. Fro...	['Joannes Rochut']	Ravenio Books	19??	['Music']	19??

	Title	Price	review/score	review/time	review/summary	review/text	description	authors	publisher	publishedDate	categories	publishedDate_year
200828	The Iliad: The Story of Achilles	21.762656	4.0	1208131200	One of our first war novels	One of our first war novels: the Achains and t...	This new, modern translation of The Iliad is f...	['Homer']	Signet Book	199?	['Epic poetry, Greek']	199?
2919171	Circular breathing for the wind performer	21.762656	4.0	1306108800	/you can find this information online	I was disappointed when I received this book. ...	Author Kynaston discusses the technique of cir...	['Trent Kynaston']	Alfred Music	199?	['Music']	199?
2919172	Circular breathing for the wind performer	21.762656	3.0	1261958400	So 10.95 for this ?	Although, of course, extremely useful, this is...	Author Kynaston discusses the technique of cir...	['Trent Kynaston']	Alfred Music	199?	['Music']	199?
2994104	The Iliad: The Story of Achilles	21.762656	5.0	927417600	The pride of Agamemnon and Achilles.	The version of the "Iliad" that I re...	This new, modern translation of The Iliad is f...	['Homer']	Signet Book	199?	['Epic poetry, Greek']	199?

	Title	Price	review/score	review/time	review/summary	review/text	description	authors	publisher	publishedDate	categories	publishedDate_year
780252	One wonderful night: A Romance of New York,	21.762656	1.0	1349654400	Not worth reading	Did not care for it. Very old fashioned and to...	Lady Hermione flees to New York to escape marr...	['Louis Tracy']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
780253	One wonderful night: A Romance of New York,	21.762656	3.0	1351468800	One Wonderful Night A Romance of New York	This book was hard to follow at times. It was ...	Lady Hermione flees to New York to escape marr...	['Louis Tracy']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
780254	One wonderful night: A Romance of New York,	21.762656	3.0	1350950400	one wonderful night a romance in new york	this was an interesting book set in a early ag...	Lady Hermione flees to New York to escape marr...	['Louis Tracy']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
1639902	Wild Sorceress	18.000000	3.0	1121644800	Story 5 stars, Writting 2 stars:	I give the story line 5 stars but the writing ...	In a world where hostile nations wield magic i...	['Margaret L. Carter', 'Leslie Roy Carter']	Writers Exchange E-Publishing	101-01-01	['Fiction']	101-
1639903	Wild Sorceress	18.000000	5.0	1313020800	A Delightful Read	This fantasy novel features a very original sy...	In a world where hostile nations wield magic i...	['Margaret L. Carter', 'Leslie Roy Carter']	Writers Exchange E-Publishing	101-01-01	['Fiction']	101-
1645910	"Charge that to my account," and other Gospel ...	21.762656	5.0	1211760000	One of the best presentations of salvation by ...	This little book shows the the gospel and salv...	We are "accepted in the beloved." The blessed ...	['H.A. Ironside']	Solid Christian Books	101-01-01	['Religion']	101-
1645911	"Charge that to my account," and other Gospel ...	21.762656	5.0	1211760000	One of the best presentations of salvation by ...	This little book shows the the gospel and salv...	We are "accepted in the beloved." The blessed ...	['H.A. Ironside']	Solid Christian Books	101-01-01	['Religion']	101-
1645912	"Charge that to my account," and other Gospel ...	21.762656	5.0	1208995200	One of the best presentations of salvation by ...	This little book shows the the gospel and salv...	We are "accepted in the beloved." The blessed ...	['H.A. Ironside']	Solid Christian Books	101-01-01	['Religion']	101-
1689407	Lore Volume 1 (v. 1)	21.762656	5.0	1148083200	Something beautifully new!	I remember when I first picked up a single Lor...	This is the first Volume of Demonology And Dev...	['Moncure Daniel Conway']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
1689408	Lore Volume 1 (v. 1)	21.762656	5.0	1138924800	Lore	I picked this book up a little while ago knowi...	This is the first Volume of Demonology And Dev...	['Moncure Daniel Conway']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
1689409	Lore Volume 1 (v. 1)	21.762656	4.0	1150070400	Poise between extremes	I mean that poise in a lot of ways. First, the...	This is the first Volume of Demonology And Dev...	['Moncure Daniel Conway']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036777	The book of talismans, amulets, and zodiacal gems	21.762656	5.0	1304467200	Excellent explanatory and directive!	This is an easy to read encyclopedia of symbol...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036778	The book of talismans, amulets, and zodiacal gems	21.762656	4.0	1347321600	a worthy addition	If you are quite like me and love all things t...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036779	The book of talismans, amulets, and zodiacal gems	21.762656	4.0	1309132800	interesting public domain book	from 1922 an amazing collection of info on sym...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036780	The book of talismans, amulets, and zodiacal gems	21.762656	3.0	1356566400	It's ok.	This book has great explanation of each sign i...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036781	The book of talismans, amulets, and zodiacal gems	21.762656	5.0	1354320000	Great!	This is a book I have been truly waiting for. ...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036782	The book of talismans, amulets, and zodiacal gems	21.762656	4.0	1359158400	Good Information	The book has lots of good information. The onl...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036783	The book of talismans, amulets, and zodiacal gems	21.762656	5.0	1356912000	Very informative	Though an older book on the subject; it is ver...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036784	The book of talismans, amulets, and zodiacal gems	21.762656	1.0	1355961600	Poorly organized and written	I found this book poorly organized and written...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036785	The book of talismans, amulets, and zodiacal gems	21.762656	2.0	1354665600	Modern books beat this one	Modern books on this subject, which I approach...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036786	The book of talismans, amulets, and zodiacal gems	21.762656	5.0	1354147200	Well done	I really have no major complaints; it's a good...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036787	The book of talismans, amulets, and zodiacal gems	21.762656	4.0	1354060800	INTERESTING	There is a lot of information in this book, I ...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036788	The book of talismans, amulets, and zodiacal gems	21.762656	4.0	1352419200	THE BOOK OF TALISMANS, AMULETS AND ZODIACAL GEMS	This book is well written. It contains insight...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036789	The book of talismans, amulets, and zodiacal gems	21.762656	3.0	1350777600	not bad, not great	I downloaded this free sample of the book of a...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2036790	The book of talismans, amulets, and zodiacal gems	21.762656	4.0	1333238400	Informative	I skimmed the information in this book. It was...	With 10 illustrative Plates. This is a study o...	['William Thomas and Kate Pavitt']	BEYOND BOOKS HUB	101-01-01	['Social Science']	101-
2069642	The Bathtub Is Overflowing but I Feel Drained:...	21.762656	5.0	1186358400	A must for all moms!	This is an excellent book for any mom looking ...	PLEASE NOTE: This is a summary and analysis of...	['ZIP Reads']	ZIP Reads	101-01-01	['Self-Help']	101-
2069643	The Bathtub Is Overflowing but I Feel Drained:...	21.762656	5.0	1259020800	Amazing!	I just finished reading this book and I loved ...	PLEASE NOTE: This is a summary and analysis of...	['ZIP Reads']	ZIP Reads	101-01-01	['Self-Help']	101-
2069644	The Bathtub Is Overflowing but I Feel Drained:...	21.762656	5.0	1223251200	LOVE THIS!!!	My husband just bought me this book after seei...	PLEASE NOTE: This is a summary and analysis of...	['ZIP Reads']	ZIP Reads	101-01-01	['Self-Help']	101-
2069645	The Bathtub Is Overflowing but I Feel Drained:...	21.762656	5.0	1215648000	Gives Mom A Quick Refreshing	I bought this book for myself, and am now purc...	PLEASE NOTE: This is a summary and analysis of...	['ZIP Reads']	ZIP Reads	101-01-01	['Self-Help']	101-
2069646	The Bathtub Is Overflowing but I Feel Drained:...	21.762656	5.0	1286928000	great book	I am reading this book a chapter at a time wit...	PLEASE NOTE: This is a summary and analysis of...	['ZIP Reads']	ZIP Reads	101-01-01	['Self-Help']	101-
2069647	The Bathtub Is Overflowing but I Feel Drained:...	21.762656	5.0	1254528000	Hilariously uplifting and informative :0)	I am absolutely in LOVE with this book! Everyt...	PLEASE NOTE: This is a summary and analysis of...	['ZIP Reads']	ZIP Reads	101-01-01	['Self-Help']	101-
2115497	His Family	21.762656	3.0	1245974400	A good heart	When you read this, it is important to conside...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115498	His Family	21.762656	5.0	1218931200	Changes in New York City life in the Early 20t...	One of my favorite Pulitzer winning pieces of ...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115499	His Family	21.762656	5.0	1357689600	love it	Great book, I would recommend it to anyone who...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115500	His Family	21.762656	1.0	1281916800	starts on chapter two	While I am sure the book is good, this copy th...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115501	His Family	21.762656	3.0	1027987200	A Nostalgic Glimpse of Life	Ernest Poole's 'His Family' focuses on the lat...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115502	His Family	21.762656	3.0	1003622400	The winner of the first Pulitzer Prize for fic...	Until coming across his name in a list of Puli...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115503	His Family	21.762656	5.0	966729600	Lovely glimpse of the past	Poole's book tells the story of aging patriarc...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115504	His Family	21.762656	3.0	1125273600	Pre-modern Modern	This novel has the distinction of winning the ...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115505	His Family	21.762656	4.0	1100995200	The Dawning of the 20th Century	This is the very first of the Pulitzer Prize w...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115506	His Family	21.762656	3.0	1201392000	The Changing Society Of The Early 1900s	"His Family" by Ernest Poole was published in ...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115507	His Family	21.762656	4.0	1359504000	Roger Gale	In 1918, Ernest Poole's novel, "His Family" be...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115508	His Family	21.762656	4.0	1303776000	Earnest and sympathetic novel	His Family is a very fine novel about widower ...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115509	His Family	21.762656	3.0	1274918400	History repeats itself	This is the first fictional book that won the ...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2115510	His Family	21.762656	4.0	1346284800	Excellent Study of the Perils of Middle Age	It could just be that the main character is ab...	His Family by American writer Ernest Poole was...	['Ernest Poole']	BEYOND BOOKS HUB	101-01-01	['Fiction']	101-
2717219	Echine (French Edition)	21.762656	5.0	962928000	Echine	My favorite book from Djian, its part of the B...	Extrait : "Avant de raconter un des plus effro...	['OCTAVE MIRBEAU']	BEYOND BOOKS HUB	101-01-01	['Literary Collections']	101-

	Title	Price	review/score	review/time	review/summary	review/text	description	authors	publisher	publishedDate	categories	publishedDate_year
1384321	A Wealth of Wisdom: Legendary African American...	11.18	5.0	1074384000	Wealth of Knowledge - History's Truths	Thanks to Dr. Cosby and Renee Pouissant we now...	The wisdom of our elders is our most valuable ...	['Camille Cosby', 'Rene Poussaint']	Simon and Schuster	2030-12-31	['Biography & Autobiography']	2030
1384322	A Wealth of Wisdom: Legendary African American...	11.18	5.0	1076544000	Couldn't Put It Down	This is one of those rare books I simply could...	The wisdom of our elders is our most valuable ...	['Camille Cosby', 'Rene Poussaint']	Simon and Schuster	2030-12-31	['Biography & Autobiography']	2030
1384323	A Wealth of Wisdom: Legendary African American...	11.18	4.0	1075939200	Enjoyed It!!	I enjoy reading this book. Really. The reason ...	The wisdom of our elders is our most valuable ...	['Camille Cosby', 'Rene Poussaint']	Simon and Schuster	2030-12-31	['Biography & Autobiography']	2030
1384324	A Wealth of Wisdom: Legendary African American...	11.18	5.0	1077494400	Great Advise	This is a wonderful book of advice from wise A...	The wisdom of our elders is our most valuable ...	['Camille Cosby', 'Rene Poussaint']	Simon and Schuster	2030-12-31	['Biography & Autobiography']	2030
1384325	A Wealth of Wisdom: Legendary African American...	11.18	5.0	1355702400	Women's Wit	So many women sharing views through their own ...	The wisdom of our elders is our most valuable ...	['Camille Cosby', 'Rene Poussaint']	Simon and Schuster	2030-12-31	['Biography & Autobiography']	2030

	Price	review/time	publishedDate_year	text length
review/score
1.0	21.653179	1.118128e+09	2004.820585	793.900548
2.0	21.711868	1.121717e+09	2005.093863	928.639748
3.0	21.673666	1.136735e+09	2005.073905	986.441841
4.0	21.678704	1.135118e+09	2004.789454	946.258080
5.0	21.556469	1.128884e+09	2004.798769	756.711431

	review/score	review/text	sentiment
1	5.0	I don't care much for Dr. Seuss but after read...	positive
2	5.0	If people become the books they read and if "t...	positive
3	4.0	Theodore Seuss Geisel (1904-1991), aka "D...	positive
4	4.0	Philip Nel - Dr. Seuss: American IconThis is b...	positive
5	4.0	"Dr. Seuss: American Icon" by Philip Nel is a ...	positive

	review/score	review/text	sentiment
1578389	5.0	wellbeing figure even noggin wisdoms heartedly...	positive
1840683	3.0	suggest gone past seemed blood I might descrip...	negative
758886	4.0	give So far draw don explain high written pret...	positive
1441591	2.0	stands 1 18 humility doesn Taoist truths perso...	negative
2564758	3.0	practically pretty paragraph Robards total AND...	negative

Goal: Demonstrate proficiency in text mining (sentiment analysis) using Python.¶

Synopsis:¶

The data package contains two files each corresponding to 212404 unique books that were available on Amazon between 1996 and 2014. 142.8 million reviews spanning May 1996 - July 2014 have also been collected. The two data files maintain the following structure:¶

Execution of required modules.¶

Task 2: Data Cleansing, Classification, and Word Clouds .¶

1. [5 Marks] Load the data into Python and perform the initial sanity checks.¶

1-1. Variable Analysis¶

Check the structure of those data and the number of null values.¶

1-2. Data Merge¶

1-3. Treating missing values¶

1-4. Treating outlier¶

2. [10 Marks] Generate initial analytical visualizations to understand the reviews (i.e., histogram) including a word cloud (using the wordcloud package). Discuss your findings.¶

2-1. Visualize what words are often included in 'review/text' using a word cloud.¶

2-2. Display of number of data per review score¶

2-3. A number of characters in each 'review/text'.¶

2-4. Visualize the number of data per publication year.¶

3. [5 Marks] Classify reviews into two ‘sentiment’ categories called positive and negative.¶

4. [10 Marks] Generate positive and negative word clouds. Discuss your findings while comparing the positive and negative summaries (you may include other graphs if needed).¶

Using Word Cloud, we visualized the high-frequency WORDS for each high and low review.¶

Task 3: Prediction¶

1. [15 Marks] Build a simple logistic regression model to predict the sentiment category based on a text- based review. Discuss your findings.¶

2. [15 Marks] Build a multinomial logistic regression model to predict the rating of a book based on its text-based review. Discuss your findings.¶

Fin.¶

2-2.　Display of number of data per review score¶